{"id":"https://openalex.org/W4398856443","doi":"https://doi.org/10.1109/iccv51701.2025.00174","title":"Towards Cross-Modal Backward-Compatible Representation Learning for Vision-Language Models","display_name":"Towards Cross-Modal Backward-Compatible Representation Learning for Vision-Language Models","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4398856443","doi":"https://doi.org/10.1109/iccv51701.2025.00174"},"language":"en","primary_location":{"id":"doi:10.1109/iccv51701.2025.00174","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.00174","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},"type":"preprint","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2405.14715","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5075209169","display_name":"Young Kyun Jang","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]},{"id":"https://openalex.org/I4210090411","display_name":"Google DeepMind (United Kingdom)","ror":"https://ror.org/00971b260","country_code":"GB","type":"company","lineage":["https://openalex.org/I4210090411","https://openalex.org/I4210128969"]}],"countries":["GB","US"],"is_corresponding":true,"raw_author_name":"Young Kyun Jang","raw_affiliation_strings":["Google DeepMind"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google DeepMind","institution_ids":["https://openalex.org/I1291425158","https://openalex.org/I4210090411"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5029216091","display_name":"Ser-nam Lim","orcid":null},"institutions":[{"id":"https://openalex.org/I106165777","display_name":"University of Central Florida","ror":"https://ror.org/036nfer12","country_code":"US","type":"education","lineage":["https://openalex.org/I106165777"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ser-nam Lim","raw_affiliation_strings":["University of Central Florida"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Central Florida","institution_ids":["https://openalex.org/I106165777"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5075209169"],"corresponding_institution_ids":["https://openalex.org/I1291425158","https://openalex.org/I4210090411"],"apc_list":null,"apc_paid":null,"fwci":1.0652,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.73218795,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":96,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"1783","last_page":"1792"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.9994999766349792,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10812","display_name":"Human Pose and Action Recognition","score":0.963699996471405,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.9617999792098999,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/modal","display_name":"Modal","score":0.7330080270767212},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.6460815072059631},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5690473914146423},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.43313926458358765},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3623361587524414},{"id":"https://openalex.org/keywords/political-science","display_name":"Political science","score":0.06626671552658081},{"id":"https://openalex.org/keywords/materials-science","display_name":"Materials science","score":0.05456435680389404}],"concepts":[{"id":"https://openalex.org/C71139939","wikidata":"https://www.wikidata.org/wiki/Q910194","display_name":"Modal","level":2,"score":0.7330080270767212},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6460815072059631},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5690473914146423},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.43313926458358765},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3623361587524414},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.06626671552658081},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.05456435680389404},{"id":"https://openalex.org/C188027245","wikidata":"https://www.wikidata.org/wiki/Q750446","display_name":"Polymer chemistry","level":1,"score":0.0},{"id":"https://openalex.org/C94625758","wikidata":"https://www.wikidata.org/wiki/Q7163","display_name":"Politics","level":2,"score":0.0},{"id":"https://openalex.org/C199539241","wikidata":"https://www.wikidata.org/wiki/Q7748","display_name":"Law","level":1,"score":0.0}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/iccv51701.2025.00174","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.00174","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2405.14715","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2405.14715","pdf_url":"https://arxiv.org/pdf/2405.14715","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2405.14715","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2405.14715","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2405.14715","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2405.14715","pdf_url":"https://arxiv.org/pdf/2405.14715","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320317153","display_name":"DeepMind","ror":"https://ror.org/00971b260"}],"has_content":{"grobid_xml":false,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4398856443.pdf"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052","https://openalex.org/W2382290278","https://openalex.org/W3204019825"],"abstract_inverted_index":{"Modern":[0],"retrieval":[1,105,186],"systems":[2],"often":[3],"struggle":[4],"with":[5,60,135],"upgrading":[6],"to":[7,14,53,73,83,91,125,196],"new":[8,23,57,122,175,202],"and":[9,22,171,193],"more":[10],"powerful":[11],"models":[12],"due":[13],"the":[15,20,36,56,61,68,74,80,103,121,128,140,157,173,189],"incompatibility":[16],"of":[17,42,70,76,127,142,191],"embeddings":[18,37,124],"between":[19,94],"old":[21,62,129,158],"models.":[24],"This":[25,65,131],"necessitates":[26],"a":[27,39,116,201],"costly":[28],"process":[29],"known":[30],"as":[31,100],"backfilling,":[32],"which":[33],"involves":[34],"re-computing":[35],"for":[38,102,146],"large":[40],"number":[41,141],"data":[43],"samples.":[44],"In":[45],"vision,":[46],"Backward-compatible":[47],"Training":[48],"(BT)":[49],"has":[50],"been":[51],"proposed":[52],"ensure":[54],"that":[55,119,168],"model":[58,159,204],"aligns":[59],"model's":[63,123,176],"embeddings.":[64],"paper":[66],"extends":[67],"concept":[69],"vision-only":[71],"BT":[72,86],"field":[75],"cross-modal":[77,104,185],"retrieval,":[78],"marking":[79],"first":[81],"attempt":[82],"address":[84,108],"Cross-modal":[85],"(XBT).":[87],"Our":[88],"goal":[89],"is":[90,152],"achieve":[92],"backward-compatibility":[93],"Vision-Language":[95],"Pretraining":[96],"(VLP)":[97],"models,":[98],"such":[99],"CLIP,":[101],"task.":[106],"To":[107],"XBT":[109,147,192],"challenges,":[110],"we":[111,163],"propose":[112],"an":[113],"efficient":[114],"solution:":[115],"projection":[117],"module":[118],"maps":[120],"those":[126],"model.":[130],"module,":[132],"pretrained":[133],"solely":[134],"text":[136],"data,":[137],"significantly":[138],"reduces":[139],"image-text":[143],"pairs":[144],"required":[145],"learning,":[148],"and,":[149],"once":[150],"it":[151,154],"pretrained,":[153],"avoids":[155],"using":[156],"during":[160],"training.":[161],"Furthermore,":[162],"utilize":[164],"parameter-efficient":[165],"training":[166],"strategies":[167],"improve":[169],"efficiency":[170],"preserve":[172],"off-the-shelf":[174],"knowledge":[177],"by":[178],"avoiding":[179],"any":[180],"modifications.":[181],"Experimental":[182],"results":[183],"on":[184],"datasets":[187],"demonstrate":[188],"effectiveness":[190],"its":[194],"potential":[195],"enable":[197],"backfill-free":[198],"upgrades":[199],"when":[200],"VLP":[203],"emerges.":[205]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-05-06T06:03:25.996018","created_date":"2025-10-10T00:00:00"}